Using Machine Learning and Google Earth Engine to Develop a High-Resolution Forest Canopy Cover Dataset: A Case Study in Arkansas, USA

The forest canopy cover is an essential factor in the health and functioning of forests, as it plays a role in ecosystem services such as carbon sequestration, wildlife habitat, and water regulation. This study developed a large-scale forest canopy cover dataset with a 1-meter spatial resolution for Arkansas, USA, using aerial imagery and a machine learning algorithm processed and analyzed with the Google Earth Engine platform. The results showed that the dataset successfully identified forest canopy cover with an overall accuracy of 83.31% and 94.35% per county. It also showed a strong positive correlation with the existing Global Forest Maps and the U.S. Forest Service products, with R
²
values ranging between 0.97 and 0.99 for the county and city levels. This dataset can monitor, forecast, and manage forest resources in Arkansas and other regions. The study also demonstrates the potential of using machine learning and cloud computing technologies to create high-resolution forest cover datasets that can be applied to other areas worldwide.